Failover Clustering in Windows Server 2016

[@Killtacboy]

This is how a lecture should be. Show the conceptual elements and then dive right in with the hands on example. Terrific.

[@adamalabtah8034]

The Amazing thing that I found in this vid, that you have grabbed the idea of the clustering and then you have mentioned something doing that in virtualization and physical environments. Appreciated MAN. Well Done!. keep sharing knowledge and keep the great work up!.

[@mihailyurukov5942]

You know the guy explaining this really understands it and is a good system administrator, when he explains it by using practical vernacular instead of academic bullshit. Excellent ! Thank you for helping me and many other get a grasp on Failover clustering.

[@BijouBakson]

Fantastic! You went a bit fast but we got the pause and rewind button; so all is well. This helped clear much. Much appreciated.

[@readerko]

Same, preparing for MCSA, amazing how you explained everything. Thank you.

[@andriuskarpas]

But what if an iSCSI server fails? Nodes in a cluster have no access to shared storage. Then this type of failover clustering makes no sense having fail domain at the storage point.

[@techytube]

You are right, the network resiliency and teaming then becomes critical. However working with networking folks in an organization you can achieve high levels of reliability for SAN connections. Still though your comment stands in that this is a failure point.

[@MrBimalnp4me]

Very clear. Thanks

[@aaweshaminsayyed1722]

Thanks a lot for sharing the knowledge in such a nice and easy way :) Cheers!!!!

[@literalbuddha2921]

Excellent explanation!

[@mustafafarhat907]

Amazing Thanks, WOuld you please explain how to ''Configuring Generic Applications, Scripts, and Services''?

[@luisvasconcelos874]

Hi, is it possible to set up 2 NAS servers on a failover cluster ? Configuring their local disks as one single storage location?

[@fatherfigure2766]

So will the private and public network be only for the nodes in the cluster? Or the Target server will have public and private connection too?

[@Devianovic2010]

hello , at 16:38 i noticed there is no storage available , but at 16:45 i noticed magicly there is a storage , how is this possible to happen so fast , i really would like to learn this new tehnique which could be very usefull , because from my knowledge you configured only one volume which is the witness disk and you are missing the available storage disk , but i noticed you managed to apply it in 7 seconds , usually it takes 5 minutes , pending your response In all it is a good demo

[@vitaliisharapov8866]

well if you will look in the computer time in the bottom right corner OP just did changes to his setup and did not record it.. just didnt bother to explain to us where was a hickup....... cuz Im watching it and its kinda confusing wtf this storage poped up from.........

[@jackedandtanned6ft5]

I have the same problem; no storage is available, how did he fix this?

[@techytube]

@@jackedandtanned6ft5 It was a transient network problem, I should have added an explanation, my apologies..

[@jeffrowe6004]

nice video...but (there is always a but) the iSCSI target is a single point of failure, how do you eliminate that? You added redundancy in the connection to the file server, but what if the file server goes down?

[@erniechiasson]

Yup lose that target you lose everything!

[@LtShifty]

In server 2016 and newer you can now use Azure Blob Storage as a FileShare witness, this eliminates the single point of failure and does not require the overhead of ensuring you have a HA file server. Obviously you need to ensure you have a resilient WAN infrastructure, but this is typically easier to maintain in most scenarios. With ISCSI you should also use Multipath with at least two NICs on separate physical networks from the nodes to the storage.

[@nik27184]

Excellent video...

[@AKSoapy29]

So is the iSCSI shared storage now your single point of failure? If that goes down, then what?

[@LtShifty]

That is correct in this example the ISCSI target is a SPOF, in an enterprise environment most organisations would use a SAN (IE HPE MSA 2052) which has at least two ISCSI/FC controllers, these would be connected two redundant physical networks, and the nodes would use at least two physical NICs to connect to the ISCSI/FC network utilising Multipath.

[@Rinke_08]

Hello, Let's imagine that we have 2 nodes cluster with quorum disk (node and disk majority, each of components has 1 vote). What happens when quorum goes down? Will cluster service be up and running? Cheers Slawek

[@olifern]

Hi Techytube, can you let me know which software was used to make this kind of a tutorial. I have a tablet and a stylus and I believe that can be used? Unless you used something else. I need it to help a relative who is not able to teach at school due to the covid lockdown

[@techytube]

Hi Oliver, The software used was TechSmith Camtasia, the tablet used was a Cintiq 13HD. Let me know if I can provide more details. There are other alternatives as well but Camtasia is one of the most convenient and easy ones to use.

[@olifern]

@@techytube Thanks!

[@000hugohew]

Hi Sir the application software need to install at the iscsi VD, so application software will run on HA version, it is correct, thank you.

[@albixhako1727]

Smooth.

[@trevorkruse449]

When I created a failover on 2012 I had to create a quorum shared drive. Is that not a requirement anymore?

[@davutyalcin5924]

my friend where r u from? and goood job.

[@EmreErdogan1979]

bana da türk gibi geldi. mükemmel anlatım için teşekkürler.

[@techytube]

@@EmreErdogan1979 gibisi fazla :)

[@apsogos4ever]

Which is the specs of the hyper visor? I saw many vms!

[@TheSanien]

Makes no sense regarding not putting vms on on host? Sure spread them out but the whole point with fail over is that is automatically migrate the VM to the whist that is working...?

[@skyltdockan]

As he said, in this demo the nodes are VMs on the same server which is bad. If the server dies then both nodes die. In production environment the nodes should be on separate physical servers.

[@colinanderson967]

He talked about that. He said it wasn't good practice, but this is just a demo.